www.elsevier.com/locate/micromeso

Microporous and Mesoporous Materials 87 (2005) 45–51

Photochemical template removal and spatial patterning of zeoliteMFI thin films using UV/ozone treatment

Qinghua Li a, Meri L. Amweg b, Chanel K. Yee b,Alexandra Navrotsky a,*, Atul N. Parikh b,*

a Thermochemistry Facility and NEAT ORU, University of California at Davis, Davis, CA 95616, USAb Department of Applied Science and NEAT ORU, University of California at Davis, Davis, CA 95616, USA

Received 9 March 2005; received in revised form 19 July 2005; accepted 20 July 2005Available online 5 October 2005

Abstract

This paper describes an application of a non-thermal, photochemical calcination process for an efficient and spatially controlledremoval of the organic structure-directing agent in the preparation of thin films of microporous or zeolite materials. We preparedthin-films of a high silica zeolite (structure code: MFI) following a previously published procedure. The films were illuminated usingan ozone generating short-wavelength ultraviolet light in ambient environments and characterized using Fourier-transform infraredspectroscopy, imaging ellipsometry, thin-film X-ray diffraction, and scanning electron microscopy. Results presented here indicatethat the UV/ozone treatment under nominally room temperature conditions leads to complete removal of template (structure-direc-ting-agent) from zeolite films comparable to that achieved by thermal calcination. Furthermore, spatially addressing the UV/ozoneillumination pattern using a physical mask resulted in the lateral confinement of the template removal from the zeolite film leavingbehind a composite film composed of templated and template-free regions. Subsequent chemical treatment of the patterned filmselectively removed the as-synthesized, unexposed, regions of the film thereby providing a means for the creation of isolated zeolitefilm islands at predetermined locations on the substrate surface.� 2005 Elsevier Inc. All rights reserved.

Keywords: Microporous thin films; Calcination; Patterning; UV/ozone; Ellipsometric imaging

1. Introduction

Thin films of microporous materials or zeolites areattracting attention because of their potential usefulnessin many technological applications including sensors,separation membranes for gases and liquids, low dielec-tric constant coatings, and as lasing media [1–5]. To thisend, several synthetic approaches have recently emergedfor the direct synthesis of zeolites on substrate surfacesas thin films. For instance, in situ crystallization [6,7]

1387-1811/$ - see front matter � 2005 Elsevier Inc. All rights reserved.doi:10.1016/j.micromeso.2005.07.048

* Corresponding authors. Tel.: +1 530 7547055; fax: +1 530 7522444(A.N. Parikh).

E-mail addresses: anavrotsky@ucdavis.edu (A. Navrotsky), anpar-ikh@ucdavis.edu (A.N. Parikh).

seeded-growth [8,9] spin-coating of colloidal zeolitecrystals [10], and vapor-transport methods [11] have allproved useful in the designed synthesis of microporousmaterials as thin films. Many of these zeolite films con-tain organic structure-directing-agents (SDA) whichneed to be removed. Most widely used template removalmethods use high-temperature thermal calcination pro-cesses employing temperatures in the 300–550 �C range.However, the use of high-temperature calcination pro-cesses can induce significant thermal stresses especiallyat the substrate–film interface, further accentuated bythe shrinkage of the zeolite framework which accompa-nies template removal. These stresses can result indefects and cracks within the film interior as well as atthe interfaces [12–14]. Such stresses can become quite

46 Q. Li et al. / Microporous and Mesoporous Materials 87 (2005) 45–51

significant for thin films. Moreover, thermal calcinationprecludes the use of temperature-intolerant and organic/polymeric materials as substrates. Previous studies haveshown that by using a carefully optimized calcinationtemperature program, crack-free thin films/membranescan be prepared, but such processes normally takes longcalcination periods, and specialized sample pre-treat-ments [15]. Thus the processing cannot be easily genera-lized. Based on the above, generally applicable lowtemperature calcination methods for template removalfrom thin films are highly desirable.Recently, we and others have shown that ozone-

generating ultraviolet illumination provides a simplelow-temperature alternative to the thermal calcinationprocess for bulk microporous [16] and mesoporousmaterials [17,18] and their thin films [19,20]. While de-tailed mechanistic understanding of the process is stillincomplete, it is generally believed that the process isphotochemical. The simplest description of the overallprocess involves three overlapping photoinduced chemi-cal reactions [16,21]. First, UV light with a wavelengthbelow 245.4 nm (optimally at k = 184 nm) facilitatesthe dissociation of oxygen (from the ambient) to pro-duce ozone and atomic oxygen. Simultaneously, the253.7 nm line emitted by the same lamp excites and/ordissociates the organic matrix, thereby producing acti-vated species, such as ions, free radicals, and excitedmolecules. Finally, the activated organic species arereadily attacked by atomic oxygen and ozone synergisti-cally to form simpler volatile (or removable) products,such as CO2, H2O, and N2 which escape the sample inte-rior. UV–ozone treatment is performed at room temper-ature under ambient air but local temperature at thesubstrate during the process often increases to 40–50 �C. By contrast, the currently used thermal calcina-tion processes employ much higher temperatures(>500 �C). Thus, this nominally room temperature pho-tochemical method may provide an efficient and simplemethod for the removal of organic SDA from zeolitesand the formation of crack-free zeolite films/mem-branes. Furthermore, non-thermal calcination offersinteresting opportunities to carry out calcination on thinfilms in a spatially controlled manner. It appears that byspatially addressing the photochemical calcination, itmay be possible to create patterns of microporous mate-rial and functionality associated with it.In the present work, high-quality silica zeolite MFI

films were synthesized following a seeded-growthmethod. The organic templates were removed by boththe thermal calcination and the photochemical, non-thermal calcination methods. The sample characteristicswere compared using a combination of X-ray diffraction(XRD), Fourier-transform infrared spectroscopy (FT-IR), imaging ellipsometry and scanning electron micro-copy (SEM) measurements. Spatial patterning wasachieved by conducting the photochemical calcination

in conjunction with a mask and subsequent gentle chemi-cal treatment was shown to enhance the patterns byselective removal of the SDA-containing regions of thepatterned MFI thin films.

2. Experimental section

2.1. Preparation of MFI thin films

A previously established seeding method [9] was usedfor the preparation of MFI zeolite films. Briefly, seedcrystals of TPA-silicalite-1 with an average size of�60 nm were first adsorbed on charge-reversed singlecrystal silicon (100) substrates. TPA-silicalite-1 (MFI)films were subsequently grown by a hydrothermal treat-ment of the above seeded silicon substrate in a precursorsolution at 100 �C for approximately 24 h. The precur-sor solution comprised the molecular constituents ina predetermined molar composition of 3TPAOH:25-SiO2:1300H2O:100EtOH where TPAOH is tetrapropy-lammonium-hydroxide and EtOH is ethanol. The silicasource was from tetraethylorthosilicate (TEOS). Subse-quently, the samples were rinsed with a 0.1 M ammoniasolution to remove sediments and any excess unreactedgel from the film surface.

2.2. Removal of SDA from as-synthesized MFI films

Each as-synthesized MFI sample was cut into twopieces to enable thermal and non-thermal calcinationon single films. One piece was exposed to ozone-genera-ting short-wavelength ultraviolet (UV) light (k = 184–257 nm), produced by a low- or medium-pressure Hgdischarge lamp (10–20 mW cm�2 UVP, Inc., Upland,CA) in a quartz envelope. The lamp is maintained in aclosed chamber under ambient laboratory conditions.The sample was placed in the UV/ozone reactor so thatthe substrates were about 2–3 mm away from the lightsource. The exposure period of �4 h was used, howeverthe exposure periods were not optimized. The secondpiece was calcined at 550 �C for 6 h using a heating rateof 0.2 �C/min and a cooling rate of 0.3 �C/min. Totalcalcination time in the thermal treatment is more than3 days.Spatially directed UV illumination was achieved by

carrying out the UV-treatment in conjunction with aphysical mask. This mask, comprising alternating fea-tures of chrome deposited on quartz, was producedusing standard photolithography techniques. The maskswere fabricated either at the Northern California Nano-technology Center at UC Davis or were purchased fromPhotoscience, Inc. (Torrance, CA). Subsequent chemicaltreatment of selected patterned samples was performedby immersing the samples in 0.2 N NaOH for fixed incu-bation periods.

Q. Li et al. / Microporous and Mesoporous Materials 87 (2005) 45–51 47

2.3. Characterization

Prior to and after UV treatment or thermal calcina-tion, XRD-data were collected with an Inel X-ray dif-fractometer (XRG 3000) using Ni-filtered CuKaradiation and operated at 30 kV and 30 mA. The filmsurface was oriented perpendicular to a plane definedby the X-ray source, sample holder and detector. Thestep size was 0.01� and the time per step was 80 s.FT-IR spectra were obtained in a reflection mode

using a IRscope II (Bruker, Gottingen, Germany) con-nected to a Bruker Equinox 55 FT-IR spectrophotome-ter (Bruker, Gottingen, Germany) equipped with aliquid nitrogen cooled MCT detector and a motorizedX–Y stage manipulator. Measurements of specific re-gions (grid area or square area) on samples were aidedwith a 3D video-assisted measurement software and aCCD camera. The mid-IR spectra in the frequencyrange of 4000–600 cm�1 were obtained at a 4 cm�1 res-olution for 500 scans using a Blackman Harris 3 termapodization. Both the microscope and the spectrometerwere purged with dry and carbon dioxide-free air. Thedata analysis was performed using Grams 32 (GalacticIndustries, Salem, NH) software.The film morphology and thickness were determined

by analysis with a Philips XL 30 SEM equipped witha LaB6 emission source. The samples for SEM weresputtered with a thin carbon layer prior to the mea-surements. Imaging ellipsometry measurements wereperformed using a commercial iElli2000 (NanofilmTechnologie, Gottingen, Germany) operating at532 nm using a frequency-doubled Nd:YAG laser(adjustable power up to 20 mW) and equipped with amotorized goniometer for an accurate selection of theincidence angle. The ellipsometer employed the typicalpolarizer–compensator–sample–analyzer (PCSA) nul-ling-configuration in which a linear polarizer (P) and aquarter-wave plate (C) yields an elliptically polarizedincident beam. Upon reflection from the sample (S),the beam is gathered via an analyzer (A) and imagedonto a CCD camera through a long working distanceCF plan 10· (NA:0.35, Nikon, Japan) objective. TheP, C, and A positions that yielded the null conditionare then converted to the ellipsometric angles, deltaand psi. Measurements were taken at an incidence angleof 60�. Thicknesses were estimated from the ellipsomet-ric angles using a computer-aided modeling programbased on Fresnel reflections. Refractive indices wereassumed to be 4.1502–0.0449i and 1.4–0.0i for siliconand the zeolite films, respectively.

3. Results and discussion

The nominally room temperature photochemical cal-cination (UV-treatment) exposes the sample films to a

freshly-cleaned UV lamp-grid with samples maintained2–3 mm below the lamp surface. The exposure time, 3–6 h, required for complete removal of the SDA and itsdecomposition products, was found to depend on thelamp power, cleanliness of the lamp, presence of addedoxygen, and the distance of the sample from the lampas well as the thickness of the zeolite film. FT-IR mea-surement shows that the 4 h exposure is adequate inour UV exposure configuration to fully remove SDAfrom MFI thin film.SEM images shown in Fig. 1(a) and (b) represent the

top- and side-view images of as-synthesized MFI filmsgrown on a seeded silicon wafer. The images reveal theformation of a continuous film displaying denselypacked crystals with a smooth surface and appears tobe free of significant cracks and pin-holes. The lateraldimensions of the crystals at the surface appears to beabout 100 nm with many crystals that have grown aslarge as 150–200 nm, suggesting fusion and growth ofinitial seed crystals during the hydrothermal treatment.The side-view images suggest the film thickness to beapproximately 440 nm. These values are in good agree-ment with previous results [17]. Corresponding SEMimages of photochemically and thermally calcined sam-ples are shown in Fig. 1(c)–(f). The images confirm thatneither calcination treatment used here gave rise to anysignificant changes in the film morphology. Subtlechanges in film thicknesses were revealed by the ellipso-metric data. These data estimate that the thickness of as-synthesized film is approximately 460 nm, while after theremoval of SDA by UV–ozone or thermal processes, thethickness of the film reduces to �420 nm. Furthermore,the SEM data did not reveal any significant cracks ordefects in the calcination treatments used in the presentstudy.The evidence for the removal of organic SDA was

furnished by FT-IR spectroscopy. FT-IR spectra ofas-synthesized and thermally calcined samples areshown in Fig. 2. The dashed traces in Fig. 2 show thehigh frequency region (2700–3100 cm�1) of the spectra.The trace for as-synthesized samples is characterizedby the presence of a number of overlapping peaks withpeak-maxima at 2973, 2966, 2939, and 2876 cm�1, whichare straightforwardly assigned to the methylene (–CH2–)and methyl (–CH3) C–H stretching vibrations [22].Lower frequency regions of the spectra could not be eas-ily resolved presumably because of low signal through-put obtained in the microscopy geometry. By contrast,in the UV–ozone treated and thermally calcined sam-ples, these peaks are noticeably absent (vanishing bellowthe noise level in our spectra), confirming the essentiallycomplete removal (residual organics well below 1% ori-ginal content) of the SDA. We further note that thespectra in Fig. 2 appear to suggest that under the condi-tions employed in the present study, NTC appearedeven more efficient for the template removal. Taken

Fig. 1. SEM micrographs of the films. Top- and side-view images of the as-synthesized film (a, b), top- and side-view images of the UV-exposed film(c, d), and top- and side-view images of the calcined film (e, f).

Wavenumber (cm-1)

27002800290030003100

Abs

orba

nce

(a.u

.)

0.00

0.01

0.02

0.03

0.04

As-synthesized

TC

NTC

287629392966

2973

Fig. 2. FT-IR traces in the high frequency mid-infrared regions (2700–3100 cm�1) for the as-synthesized and the UV-exposed (square) MFIthin film.

5 10 15 20 25 30 35 40

0

5000

10000

15000

20000

(c)

(b)

(a)

Inte

nsity

2-Theta (o)

Fig. 3. XRD patterns of the as-synthesized film (a), the UV-exposedfilm (b) and the calcined film (c).

48 Q. Li et al. / Microporous and Mesoporous Materials 87 (2005) 45–51

together, these results are in excellent correspondencewith our earlier results wherein MFI bulk powder wasshown to lose all measurable template upon photo-and thermal-calcinations [16].Additional evidence for the removal of SDA by UV

exposure is provided by thin film XRD. The diffracto-grams shown in Fig. 3 (panels a–c) represent the XRD

patterns obtained for the as-synthesized film, the UV–ozone treated film and the thermally calcined film,respectively. Strong XRD peaks of the supporting layerat 56.2� ((001) oriented silicon wafer) in the supportedMFI zeolite film (not shown in the figure) and randomlyoriented pattern indicate that the coated MFI zeolitefilm is thin, as confirmed by SEM images. The XRDpattern of the as-synthesized thin film shows that the

Fig. 4. Optical microscope image of patterned thin film (a) showingcontrast between UV-exposed region (brighter square) and UV-unexposed region (darker grid). The square features were 25 lm inlinear dimensions. (b) FT-IR traces in the high frequency mid-infraredregions (2700–3100 cm�1) for the UV-exposed (square) and UV-protected (grid) regions of a photochemically patterned MFI thin film.

Q. Li et al. / Microporous and Mesoporous Materials 87 (2005) 45–51 49

intensities of the first two lines at about 7.9� and 8.8� 2hare lowered due to the presence of SDA in the intracrys-talline voids (see Fig. 3(a)). When the sample is treatedby UV exposure or thermal calcinations, the relativeintensity changes, i.e., the first two lines relatively in-creased in intensity whereas the lines at about 11.9�and 12.5� 2h decreased in intensity, as shown inFig. 3(b) and (c). These intensity changes can be ex-plained as the expected consequence of a decrease inloading, i.e., removal of extra framework organic(SDA) and inorganic species incorporated into the struc-tural voids during synthesis [23]. Note also that theintensity increase for the NTC samples in Fig. 3 aremore pronounced than for TC samples, further confir-ming that under the conditions employed, NTCappeared more efficient for surfactant removal than TC.Taken together, the experimental evidence above

confirms that the photochemical calcination using ozonegenerating UV light provides an alternative to thermalcalcination for zeolitic thin films. Our results suggestthat all features of the thermal calcination (1) near-com-plete template removal, (2) the retention of the inorganicframework, and (3) film morphology attributes areachieved using the nominally room temperature photo-chemical calcination process.Conventionally, zeolite films must be carefully treated

with temperature-programmed calcination process toavoid the formation of crack, especially for thin films.The total calcination time of thin film used in such treat-ments is often more than 3 days, under high tempera-ture (550 �C) conditions. By contrast, the UV-treatedmethod occurs at low temperature (<50 �C) conditions,and the total calcination time for the removal of SDA isdramatically shortened (from 3 days to 4 h). This mayprovide a more efficient and economical way to preparehigh-quality zeolite films/membranes. Crack-free thinMFI zeolite membranes treated by UV exposure willbe applied in permeation and flux tests in the future.An attractive feature of the UV–ozone template re-

moval method is the ability to pattern zeolite thin films,which has many potential applications in sensor micro-arrays, catalyst screening, and nanochemistry. Patternsas small as 25 lm was used to mask as-synthesizedMFI thin film. After UV exposure for 4 h, optical pat-terns are easily observed using an imaging ellipsometer,as shown in Fig. 4(a). The brighter square region indi-cates regions of the sample that received direct UVexposure, while the darker grid region reflects samplesareas that remained protected from the UV illumina-tion. Optical contrast derives from the thickness de-crease and/or index of refraction change between theexposed and unexposed region due to template removaland/or structure condensation. Fig. 4(b) shows FT-IRspectra of UV-exposed (square) and UV-unexposed(grid) areas. These data show that the characteristicpeaks in the CH stretching frequency range (2800–

3000 cm�1) are readily observed in the grid areas, whilethese peaks are absent in the square regions of the sam-ple. These results indicate that the template is removedselectively only from the UV exposed parts of the sam-ple, confirming the patternability of the photochemicaltemplate removal process.Following the UV exposure, subsequent chemical

processing of the films can be used to enhance theobserved patterns through preferential etching of theUV-unexposed (grid) regions of the samples. The pat-terned film was immersed in an aqueous 0.2 N NaOHsolution. At various times the sample was removed fromthe solution and examined for the variations in opticalcontrasts using imaging ellipsometry. These data areshown in Fig. 5. With the increase in exposure time to

Fig. 5. Ellipsometric images of patterned films immersed in a 0.2 N NaOH solution for three different incubation periods: (a) 10 min, (b) 30 min,(c) 80 min.

50 Q. Li et al. / Microporous and Mesoporous Materials 87 (2005) 45–51

the NaOH solution, the optical contrast gradually re-versed as the thickness of the UV unexposed regionsdiminished from above that of the exposed parts tonearly zero. This indicates the grid part is preferentiallyremoved relative to the square one. Thus, patternedislands of MFI thin films result as an end point in whichthe as-synthesized film region containing SDA iscompletely removed, while the UV-exposed film regionwithout SDA remains. More details are being investi-gated in our laboratories.

4. Conclusion

In summary, a nominally room temperature, UV/ozone treated method is shown to provide a useful alter-native to the high-temperature thermal calcinationmethods. This method provides an efficient and rapidapproach for the removal of template from zeolite thinfilms under near room temperature conditions whileretaining the inorganic framework. Because of low tem-perature used in the photochemical process, the forma-tion of cracks and defects, such as may be caused bythe different expansion between the zeolite and the sup-port in thin films during thermal calcination process, canbe minimized, thereby allowing the design of low defectfilms and membranes. In addition, the photochemicalcalcination offers a new patterning capability as demon-strated here. Spatially directed UV exposure on zeolitethin films for selective removal of template from filmswas shown to result in template-loaded and template-free microporous regions within a single contiguous thinfilm sample. We envisage such composite structures tobe useful to direct microporosity associated functional-ity to specific regions of the sample and allow directcomparison of the roles of templates on zeolite filmproperties. Moreover, UV-patterned zeolite thin filmswhen subjected to subsequent simple chemical treat-ments was found to allow the creation of patterned is-lands of zeolite thin films. The method appears generaland may be useful in the design of multifunctional elec-

tronic, optical, catalytic and sensing platforms usingmicroporous thin films.

Acknowledgment

This work received support from the NationalScience Foundation (Grant #DMR01-01391). A.N.P.,M.L.A., and C.K.Y. acknowledges support through agrant from the Office of Science, US Department of En-ergy (DE-FG02-04ER46173).

References

[1] M.E. Davis, Nature 417 (2003) 813.[2] K.J. Balkus Jr., L.J. Scottile, S.J. Riley, B.E. Gnade, Thin SolidFilms 260 (1995) 4.

[3] T. Bein, Chem. Mater. 8 (1996) 1636.[4] N. Nishiyama, K. Ueyama, M. Matsukata, Micropor. Mater. 7(1996) 299.

[5] Z. Shan, W.E.J. van Kooten, O.L. Oudshoorn, J.C. Jansen, H.van Bekkum, Micropor. Mesopor. Mater. 34 (2000) 23.

[6] E.R. Geus, M.J. den Exter, H.J. van Bekkum, Chem. Soc.,Faraday Trans. 88 (1992) 3101.

[7] Z. Wang, Y. Yan, Chem. Mater. 13 (2001) 1101.[8] L. Boudreau, M. Tsapatsis, Chem. Mater. 9 (1997) 1705.[9] J. Hedlund, S. Mintova, J. Sterte, Micropor. Mesopor. Mater. 28(1999) 185.

[10] S. Mintova, T. Bein, Adv. Mater. 13 (2001) 1880.[11] E. Kikuchi, K. Yamashita, S. Hiromoto, K. Ueyama, M.

Matsukata, Micropor. Mater. 11 (1997) 107.[12] K. Wegner, J. Dong, Y.S. Lin, J. Membr. Sci. 158 (1999) 17.[13] E.R. Geus, H. van Bekkum, Zeolites 15 (1995) 333.[14] J. Dong, Y.S. Lin, M. Hu, R.A. Peascoe, E. Andrew Payzant,

Micropor. Mesopor. Mater. 34 (2000) 241.[15] J. Hedlund, J. Sterte, M. Anthonis, A. Bons, B. Carstensen, N.

Corcoran, D. Cox, H. Deckman, W. Gijnst, P. de Moor, F. Lai, J.McHenry, W. Mortier, J. Reinoso, J. Peters, Micropor. Mesopor.Mater. 52 (2002) 179.

[16] A.N. Parikh, A. Navrotsky, Q. Li, C.K. Yee, M.L. Amweg,Micropor. Mesopor. Mater. 76 (2004) 17.

[17] M.T.J. Keene, R. Denoyel, P.L. Llwewllyn, Chem. Commun.(1998) 2203.

[18] M.T.J. Keene, R.D.M. Gougeon, R. Denoyel, R.K. Harris, J.Rouquerol, P.L.J. Llwewllyn, Mater. Chem. 9 (1999) 2843.

Q. Li et al. / Microporous and Mesoporous Materials 87 (2005) 45–51 51

[19] T. Clark, J.D. Ruiz, H. Fan, J. Brinker, B.I. Swanson, A.N.Parikh, Chem. Mater. 12 (2000) 3879.

[20] A.M. Dattelbaum, M.L. Amweg, L.E. Ecke, C.K. Yee, A.P.Shreve, A.N. Parikh, NanoLetters 3 (2003) 719.

[21] J.R. Vig, J. Vac. Sci. Technol. A 3 (1985) 1027.

[22] L.J. Bellamy, The Infrared Spectra of Complex Molecules,Chapman and Hall, London, 1975, pp. 374–383.

[23] E.L. Wu, S.L. Lawton, D.H. Oison, A.C. Rohrman Jr., G.T.Kokotailo, J. Phys. Chem. 83 (21) (1979) 2777.